JPH11249089A - Electrically controllable optical attenuator - Google Patents

Abstract

PROBLEM TO BE SOLVED: To form a Mach-Zehnder waveguide structure having a variable reflectivity element in one arm for changing a wave spectrum characteristic of a waveguide network so as to provide a specific attenuation at a specific wavelength. In a conventional electrically controlled optical attenuator, the attenuation depends on the wavelength because of spectral characteristics. According to the present invention, a Mach-Zehnder network having an electrically controllable optical path length adjuster in a longer interference arm and an electrically controllable optical path length adjuster in a shorter interference arm are provided. The wavelength dependence of the attenuator is reduced by the series connection with the Mach-Zehnder network.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an electrically controllable optical attenuator.

[0002]

BACKGROUND OF THE INVENTION British Patent Application GB 2 187 858
A discloses an electrically controllable single mode optical fiber branching element having a variable optical output ratio constituted by a series arrangement of two 4-port 3 dB quartz fiber couplers having a Mach Zehnder structure. . In a device of this kind schematically illustrated in FIG. 1, a first 4-port 3 dB single-mode quartz fiber coupler 10 with ports 10a, 10b, 10c and 10d and a coupling area 10e comprises: Using the full length of the two single mode fibers, ports 11a, 11
b, 11c and 11d, and the coupling region 11
e is optically coupled to a second four-port 3 dB single-mode quartz fiber coupler 11 with e. Between port 10b and port 11a, and between port 10c and port 11d
The total length of the single mode optical fibers 12 extending between
And 13 constitute two interference arms of a Mach-Zehnder structure. Therefore, when the optical path lengths of the interference arms 12 and 13 match, the port 10 of this structure is used.
All the light incident on port a emerges from port 11c, and all the light incident on port 10d emerges from port 1b. The optional elastic optical path length of a waveguide through which light propagates is the product of the physical length and the effective refractive index of the light propagating in the waveguide. If the optical path lengths of the two arms do not match, the light incident on port 10a is distributed between port 11b and port 11c at a ratio determined by the phase difference introduced by the difference in optical path length. By increasing the difference in optical path lengths for a given wavelength, the proportion of light reaching port 11c from port 10a varies according to a rising cosine characteristic. If the power emerging from port 10a via port 11b is absorbed or processed, the optical coupling between port 10a and port 10b can be considered a structure that operates as an optical attenuator. By introducing some type of electrically biased optical path length adjuster 14 in one of the interference arms, port 10a and port 1
The optical coupling during 1b can be considered as a structure that operates as an electrically controllable optical attenuator. UK patent application GB
2 In the case of the specific structure described in 187858A,
The adjuster is an electrostriction adjuster that changes the optical path length by physically expanding or contracting one interference arm fiber.

[0003] The attenuation value provided by the device is determined by the phase difference introduced by the difference in the optical path lengths of the two interference arms, so that the attenuation value provided by the structure does not depend on the wavelength. I can't. For some applications, this wavelength dependence is small enough to be acceptable, while for others it is unacceptably large. For a structure with interference arms of equal optical path length under zero bias conditions, FIG. 2 illustrates the various effects caused by locally increasing the effective refractive index of a 1 mm long section in one interference arm by heating. For the quantity representing the optical path length imbalance of
FIG. 4 is a graph showing the variation in attenuation calculated over the (free space) wavelength range from 1530 nm to 1560 nm. As shown in the figure, in the entire wavelength range, the spectral change of the attenuation is very small in proportion to the attenuation up to about 5 dB, but the spectral change of the attenuation is very small for the attenuation of about 15 dB or more. It turns out to be serious.

[0004]

SUMMARY OF THE INVENTION It is an object of the present invention to provide an electronic control with reduced wavelength sensitivity compared to a conventional Mach-Zehnder type structure attenuator, for example as described in GB 2 187 858A. To provide a possible optical attenuator.

[0005]

SUMMARY OF THE INVENTION The object of the invention is achieved by using a series arrangement of two Mach-Zehnder structures having interference arms of unequal optical path length under zero bias conditions. By arranging the optical path length adjuster of one Mach-Zehnder structure on the shorter interference arm and arranging the optical path length adjuster of the other Mach-Zehnder structure on the longer interference arm, If all the regulators operate with the same sensitivity (ie, both regulators operate to increase the optical path length, or both regulators operate to decrease the optical path length), It is possible to at least partially offset the wavelength sensitivity of one structure from the wavelength sensitivity of the other structure.

[0006] Other features and advantages of the present invention will be readily understood from the following description of preferred embodiments thereof, taken in conjunction with the accompanying drawings.

[0007]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS Referring to FIG.
The attenuator of this embodiment includes two Mach-Zehnder optical waveguide networks 30a and 30b optically connected in series. The optical waveguide networks 30a and 30b include an input waveguide 31a.
And 31b and a first 4-port type 3 dB coupler 33
a and 33b, waveguides 34a, 35a and 34b, 35b including two optically parallel interference arms, and
Output waveguides 32a and 32 optically coupled by series coupling of second 4-port 3dB couplers 36a and 36b
b. The interference arm 34a is longer than the interference arm 35a. Similarly, the interference arm 34b is
Longer than. The electrically controllable optical path length extension adjuster 37a is provided on the longer interference arm 34a of the waveguide network 30a, and similarly, the electrically controllable optical path length extension adjuster 37b is connected to the shorter interference arm 35b of the waveguide network 30b. Is provided.

[0008] The relative optical path length of the interference arms 34a and 35a, under zero bias conditions of the adjuster 37a, the second 3d
It is determined that substantially 100% coupling can be obtained between the input waveguide 31b and the output waveguide 32b on the B coupler 36a side. A 4-port 3 dB coupler splits light incident on one port into two equal amplitude components. Thus, referring to the coupler 10 of FIG. 1, the light incident on the port 10a is split into two components by the coupling region 10e, and the two components emerging from the coupling region 10e are respectively applied to the ports 10b and 10c. Propagate. Component 10b is referred to as a "straight-forward" component because it propagates at the same wavelength as when the light was first incident. In contrast, the other component is called the "crossover" component. As a general physical characteristic of such a 4-port 3 dB coupler, two components whose crossing component is advanced by π / 2 with respect to the linear component appear from the coupling region. Thus, for each waveguide network 30a and 30b, the required 100% coupling between the input and output waveguides is that n is a positive integer and the optical path length difference between the two arms is 2nπ. This is relevant when dealing with a phase difference.
The reason is that the input waveguide and the output waveguide do not form part of the same waveguide in any of the waveguide networks. On the other hand, in any of the waveguide networks, if the output waveguide is connected to another port, the input waveguide and the output waveguide become a part of the same waveguide.
Is given by the difference in the optical path length of the interference arm corresponding to the phase difference of (2n-1) π.

FIGS. 4 and 5 show that, for waveguide networks 30a and 30b, respectively, for different values of the bias giving equal increments of the difference in optical path length, from 1530 nm to 1560 nm.
The variation in attenuation calculated over the free space wavelength range down to nm is shown. In each case, the interference arm optical path difference under zero bias conditions corresponds to a phase angle difference of 2π at 1545 nm (free space wavelength). 2 and 4, it can be seen that the waveguide network 30a of FIG. 3 shows the same type of wavelength sensitivity as the waveguide network of FIG. 1, but the wavelength dependence of the waveguide network 30a is stronger. Comparing FIG. 2, FIG. 4 and FIG. 5 similarly, the waveguide network 30b shows an intermediate wavelength sensitivity between the wavelength sensitivity of the waveguide network of FIG. 1 and the wavelength sensitivity of the waveguide network 30a of FIG. It can be seen that the sign of the sensitivity is inverted (more attenuation is obtained at the longer wavelength, not at the shorter wavelength).

Therefore, in the case of the series arrangement shown in FIG. 3, the wavelength sensitivity of one of the waveguide networks 30a and 30b acts to offset the wavelength sensitivity of the other waveguide network. By judiciously weighting the optically generated optical path differences, the wavelength sensitivity of the series arrangement can be minimized. To achieve this weighting, for example, one of the controllers 37a and 37b is made longer than the other,
As a result, by applying the same level of bias to both regulators, the phase difference of the interference arm with the longer regulator is made smaller than the phase difference of the interference arm with the shorter regulator. You just need to increase it. Another weighting device weights by applying a proportionally larger bias to one regulator than the other. Such a proportional relationship is realized, for example, by using a voltage divider (not shown). Each waveguide network 30a of FIG.
And 30b, when using a phase difference of 2π under zero bias conditions, the greater wavelength sensitivity exhibited by the waveguide network 30a is due to the minimum wavelength dependence obtained by the series arrangement of both waveguide networks. This means that the optimal weights are in the vicinity of ratios 1 to 4. Very even weighting is achieved by the two waveguide networks 30a and 30a.
This is realized by using various values of the difference of the zero bias optical path length of the interference arm of b. For example, the difference in optical path length is
A phase difference of 2π may be given to the waveguide network 30a, and a phase difference of 4π may be given to the waveguide network 30b. Waveguide network 3
On the other hand, by switching the connection between the input waveguide and the output waveguide such that the 100% coupling is “straight” instead of “crossover” on the other hand, the phase difference between the two waveguide networks is reduced. The difference is reduced from 2π to π. For example, if waveguide network 30a maintains a "crossed" network, while waveguide network 30b is converted to a "straight-through" network, a 2π phase difference is utilized in waveguide network 30a to derive a 3π phase difference. Wave network 3
0b will be available.

FIG. 6 shows a schematic arrangement for a series arrangement of waveguide networks 30a and 30b that gives various amounts of optical path length imbalance when the zero bias phase difference is 2π in each case (at 1545 nm). For different optimally weighted values (1 to 4), the variation in attenuation calculated over the free space wavelength range from 1530 nm to 1560 nm is shown. It should be particularly noted that the wavelength sensitivity of such a series arrangement is significantly reduced from the wavelength sensitivity calculated for the single waveguide network of FIG.

A 3-port 3 dB Y-shaped coupler is:
It can be used instead of the 4-port 3 dB coupler of the series network arrangement of FIG. FIG. 7 is a diagram showing the series arrangement of the waveguide networks 70a and 70b. The difference between the series arrangement shown in FIG. 3 and the direct arrangement shown in FIG. Formula 3 dB couplers 33a, 33b and 36a, 36b are first and second three-port formulas 3, respectively.
dB Y-shaped couplers 73a, 73b and 76a, 7
6b. By using a 3 dB Y-shaped coupler, the power incident on the common arm is evenly distributed between the two branch arms,
There is no phase difference between the components incident on the branch arm. Therefore, the input waveguide 31a and the output waveguide 32a
In order to obtain 100% coupling between the two interference arms 34a and 35a, the difference in optical path length needs to coincide with the phase of 2nπ. The same relationship is required for waveguide 70b. When the difference in optical path length corresponds to the difference in phase angle of (2n-1) π, optical power does not enter in the single waveguide core mode of the waveguide 32a. Instead,
Power propagates in first-order and higher (non-guiding) modes. 1
The higher order modes are inherently much more attenuated than the single waveguide core mode. However, if the spacing between the Y-couplers 76a and 71b is so small that this natural damping is not sufficient, some form of cladding mode stripper 78 may be used to increase the damping.

As described above, in the case of the waveguide networks 30a and 30b shown in FIG. 3 using the 4-port type 3 dB coupler, whether the input waveguide and the output waveguide are arranged in a "crossing" or "straight-forward" structure. 2nπ or (2n−
1) A 100% coupling can be provided between an input waveguide and an output waveguide of a waveguide network by a difference in optical path length corresponding to a phase angle of π. Waveguide networks 70a and 70
If there is only one possible configuration for b, the only phase angle option for 100% coupling is a phase angle of 2π. However, various phase angle relationships for 100% coupling are obtained by the hybrid Mach-Zener waveguide networks 80 and 90 shown in FIGS. 8 and 9, respectively. The waveguide network 80 of FIG. 8 has an electrically controllable optical path length adjuster 87 provided on the longer interference arm 84 as shown by the solid line in FIG. 8 or a shorter as shown by the dashed line. It is similar to the waveguide network 30a or 30b of FIG. 3, depending on whether it is provided on the other interference arm 85. In the waveguide network 80 of FIG. 8, the one 4-port 3 dB coupler 86 is unchanged, while the other 4-port 3 dB
This is basically different from the waveguide network of FIG. 3 in that the coupler 33a is replaced by a 3 dB Y-shaped coupler 83. The waveguide network 90 shown in FIG. 9 is a 4-port 3 dB coupler 83.
Is replaced by a 3-port 3 dB coupler 93, and 3
8 in that the port type 3 dB coupler 86 is replaced by a 4-port type 3 dB coupler 96.
Different from 0. In the waveguide network 80, the input is port 8
When supplied to 3a, the component in interference arm 85 leads the component in interference arm 84 by π / 2. In such a situation, the zero-bias 100% coupling condition is that the optical path length of the interference arm 84 is zero and the interference arm 8
This is provided when the optical path length is longer than the optical path length of No. 5 by an amount corresponding to the phase angle (2n + ／) π. Correspondingly, if an input is provided to port 83d, a phase angle of (2n-1 / 2) π provides the required coupling.

[0014]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS Graphs of FIGS. 10, 11 and 12 are generated by repeating the same modeling as that of generating the graphs of FIGS. 4, 5 and 6 with respect to a Mach-Zehnder network using a 3-port Y-coupler. Is done. FIGS. 10 and 11 show the 3-port 3 dB of FIG. 7, respectively.
This is a specific example of a coupler-type waveguide network. For waveguide networks 30a and 30b having a zero-bias interference arm length difference corresponding to a phase difference of 2π, the calculated wavelength dependence is changed to various applied bias levels. It is a graph shown with respect to FIG. FIG. 12 is also a graph showing the wavelength dependence for a series arrangement of waveguide networks 30a and 30b, also operated with a weighting ratio of 1: 3. FIGS. 10 to 12 are diagrams obtained by measurements on the embodiment of FIG. 7, which is a series arrangement of Y-shaped coupler waveguide networks 30a and 30b, each having a difference in zero bias interference arm length corresponding to a phase difference of 2π. 13 provides a basis for comparison with the graphs shown in FIGS. Referring to FIGS. 12 and 15, the comparison is not a direct comparison as long as the weighting ratio for the graph of FIG. 15 is 1: 1. The Y-coupler network includes coating a silicon dioxide buffer layer on a planar silicon substrate, coating the silicon dioxide doped core glass layer on the buffer layer, and then patterning,
1 is an integrated optical channel waveguide structure fabricated using silicon dioxide-silicon technology that coats a silicon dioxide doped cladding glass layer. At each stage, the coating is plasma enhanced chemical vapor deposition (PE)
CVD). The core glass layer has a refractive index of 7 × 1 than the material of the buffer layer and the cladding layer.
0 ^-3 just germanium doping concentration to be increased is made by silicon dioxide doped. The waveguide of the waveguide network has a cross section of 6 μm width and 5 μm height. The curvature of the S-shaped bend of the Y-shaped coupler is 30 mm, and the distance between the two interference arms of the waveguide network near the optical path length adjuster is 160 μm. Each controller has a length of 2 mm and a width of 12
A Joule effect heating element created by patterning a sputtered layer of chromium so as to obtain separate heaters of μm size.

[Brief description of the drawings]

FIG. 1 is a schematic diagram of a Mach-Zehnder structure of a prior art optical attenuator.

FIG. 2 is a graph showing calculated spectral characteristics of a specific example of a prior art optical attenuator.

FIG. 3 is a schematic diagram of an attenuator according to a first embodiment of the present invention.

FIG. 4 is a graph showing calculated spectral characteristics of a specific example of a Mach-Zehnder optical waveguide network having an optical path length adjuster in a longer interference arm.

FIG. 5 is a graph showing calculated spectral characteristics of an example of a Mach-Zehnder optical waveguide network having an optical path length adjuster in a shorter interference arm.

FIG. 6 is a graph showing calculated spectral characteristics of a specific example of the attenuator shown in FIG. 3;

FIG. 7 is a schematic diagram of an attenuator according to a second embodiment of the present invention.

FIG. 8 is a schematic view showing another type of the hybrid Mach-Zehnder structure.

FIG. 9 is a schematic view showing another type of the hybrid Mach-Zehnder structure.

FIG. 10 shows an example of a Mach-Zehnder structure constructed using a 3-port 3-dB Y-shaped coupler instead of a 4-port 3-dB coupler; 9 is a graph showing calculated spectral characteristics of a specific example of the zender structure optical waveguide network.

FIG. 11 relates to an example of a Mach-Zehnder structure constructed using a 3-port 3-dB Y-shaped coupler instead of a 4-port 3-dB coupler; 9 is a graph showing calculated spectral characteristics of a specific example of the zender structure optical waveguide network.

FIG. 12 shows calculated spectral characteristics of one specific example of an attenuator for an example of a Mach-Zehnder structure constructed using a 3-port 3-dB Y-coupler instead of a 4-port 3-dB coupler. It is a graph.

FIG. 13 relates to an example of a Mach-Zehnder structure constructed using a 3-port 3-dB Y-shaped coupler instead of a 4-port 3-dB coupler; 5 is a graph showing measured spectral characteristics of a specific example of the zender structure optical waveguide network.

FIG. 14 relates to an example of a Mach-Zehnder structure constructed using a 3-port 3-dB Y-shaped coupler instead of a 4-port 3-dB coupler; 5 is a graph showing measured spectral characteristics of a specific example of the zender structure optical waveguide network.

FIG. 15 shows measured spectral characteristics of one specific example of an attenuator for an example of a Mach-Zehnder structure constructed using a 3-port 3-dB Y-coupler instead of a 4-port 3-dB coupler. It is a graph.

[Explanation of symbols]

 30a, 30b Mach-Zehnder optical waveguide network 31a, 31b Input waveguide 32a, 32b Output waveguide 33a, 33b First 4-port type 3dB coupler 34a, 35a, 34b, 35b Interference arm 36a, 36b Second 4-port 3dB Coupler 37a, 37b Electrically Controllable Optical Path Length Extension Controller

 ──────────────────────────────────────────────────の Continuing on the front page (72) Inventor Stephen Day Essex CM 186 A.B., Harlow, Westfield 18 (72) Inventor Terry Victor Clap Hartfordshire S.G.E. , Green Reyes Cottages 7

Claims (3)

[Claims] A first Mach-Zehnder optical waveguide network optically coupled in series, each Mach-Zehnder optical waveguide network comprising: an input waveguide; a first 3 dB coupler; A Mach-Zehnder optical waveguide network, comprising: two optically parallel interference arm waveguides; and an output waveguide optically coupled to the input waveguide by serial coupling of a second 3 dB coupler. One of the interference arms is provided with an optically biased optical path length adjuster. Under the zero bias condition of the adjuster, the two interference arms of each of the Mach-Zehnder optical waveguide networks have: Substantially 100% between the input waveguide and the output waveguide
There is a non-zero amount of optical path length difference that gives an interference condition to the second 3 dB coupler of the Mach-Zehnder optical waveguide network that provides the Mach-Zehnder optical waveguide. An electrically controllable optical attenuator provided in a longer interference arm waveguide of the waveguide network and a shorter interference arm waveguide of the other Mach-Zehnder optical waveguide network. 2. An electrically controllable optical attenuator according to claim 1, wherein said controller comprises an electric heating element. 3. The optical waveguide of each of the first and second Mach-Zehnder structures under zero bias conditions of the regulator of the optical network of each of the Mach-Zehnder structures. 2. The electrically controllable light according to claim 1, wherein the two interference arm waveguides of the net have a difference in optical path length by a non-zero amount that gives a phase angle difference of nπ when n represents a positive integer. Attenuator.